Recent attachment-related studies have demonstrated that both childhood behavior problems and adolescent psychopathology are predicted by disorganized infant attachment behavior, behavior that is characterized by conflicting behavioral tendencies and the lack of a coherent relational strategy for dealing with stress. However, based on current literature, it is unclear whether a validated measure of disorganized attachment in adolescence exists. The first aim of the proposed study is to develop and validate a coding protocol for identifying controlling-punitive, controlling-caregiving, and other insecure-disorganized behavior in adolescence. The coding scheme will be based on previous work in the field and will be applied to two attachment-related parent-adolescent interaction assessments. Participants will be 120 adolescents and their mothers from low-income families, 65 of whom who have participated in a longitudinal study at ages 12 and 18 months, 4-5 years, and 7-9 years. The construct validity of the new measure of controlling attachment strategies will be assessed in relation to coding of Unresolved or Cannot Classify attachment strategies as assessed by the Adult Attachment Interview and will also be validated against broader aspects of parent-adolescent interaction assessed in a standard revealed differences conflict resolution task, as coded by the Autonomy and Relatedness Scales. The second aim of the study is to assess whether overall risk in infancy is an important antecedent of disorganized/controlling attachment strategies in adolescence. Mediational models will test whether the onset of behavior problems in the early school years or the mother's lack of facilitation of automony and relatedness in adolescence adds to and/or mediates any observed relation between early relational risk and adolescent attachment behaviors. The third aim of the study is to assess the degree to which adolescent disorganized/controlling attachment strategies are associated with adolescent psychiatric morbidity. Psychiatric diagnoses will be assessed by the Structured Clinical Interview for Diagnosis (SCID) Axis I, the borderline and antisocial personality disorder sections of the SCID II, the Center for Epidemiologic Studies Depression Scale (CES-D) , and the Adolescent Dissociative Experiences Scale (ADES). Longitudinal analyses will further assess the degree to which early relational risk and early school age behavior problems are important precursors of adolescent psychopathology. The proposed study will contribute to increased understanding of long-term developmental trajectories that eventuate in psychopathology. In order to implement prevention or treatment programs for reducing adolescent antisocial behavior and psychopathology, it is essential to seek a thorough understanding of the developmental pathways through which such behavior develops over time.